research Market Intelligence /marketintelligence/en/news-insights/research/smart-speakers-take-off content

Login to Market Intelligence Platform

New User / Forgot Password

Looking for more?

Contact Us

Request a Demo

You're one step closer to unlocking our suite of comprehensive and robust tools.

Fill out the form so we can connect you to the right person.

  • First Name*
  • Last Name*
  • Business Email *
  • Phone *
  • Company Name *
  • City *

* Required

In this list

Smart Speakers Take Off

C&I Loan Growth Pops In Q2, But Tax Reform’s Role Remains Unclear


StreetTalk Episode 27: Looking For The Cream Of The Crop In Bank Stocks

Loans And Deposits Continue Uphill Climb At US Banks In June

Peeking Into The Future Without Staring At A Crystal Ball: Brexit Scenarios And Their Impact On Energy Firms’ Credit Risk

Technology, Media & Telecommunications
Smart Speakers Take Off


The following post comes from Kagan, a research group within S&P Global Market Intelligence.

To learn more about our TMT (Technology, Media & Telecommunications) products and/or research, please request a demo.

Jun. 25 2018 — As the smart home market continues to develop in fits and starts, one device has definitely managed to catch the attention of consumers: the smart speaker.

Smart speakers, also called virtual assistants or smart home speakers, have become increasingly popular over the past eighteen months, with total unit shipments surpassing 31 million in 2017, up from just 6.5 million in 2016

Smart Speaker worldwide unit shipments, 2015 - 2017

While smart speakers can function like a traditional speaker, allowing users to play music from a smartphone or the cloud, they also offer myriad other capabilities. One of their most popular features is the ability to respond to voice commands. This allows users to receive hands-free information about the weather, local traffic or the latest sports scores. Voice commands can also do things such as turn on a coffee maker, adjust the thermostat, read an audio book, and even change the channel on the TV.

All smart speakers rely on artificial intelligence, or AI, to carry out tasks, although the responsiveness and raw intelligence offered by smart speakers varies by product. A basic function of AI in a smart speaker is that it permits the smart speaker to learn about a user’s habits and preferences over time, thus allowing the speaker to refine and improve its search responses over time.
Two vendors, Amazon with its Echo products and Google with its Google Home smart speaker, have been the leading proponents – and the largest sellers – of smart speaker products. Other notable products include:
• Apple’s HomePod
• Sonos’ One smart speaker
• Alibaba’s Tmall Genie X1
• Xiaomi’s Mi AI smart speaker

In addition, several other vendors such as LG Electronics,, Naver, Baidu and Harmon Kardon offer smart speaker products. Even Facebook recently announced that it was planning to launch a smart speaker product before the end of 2018.

Voice assistants
Voice assistants are generally viewed as the key functionality feature in a smart speaker. For many consumers considering a voice assistants platform, they often also have to select an ecosystem, as the leading voice assistants do not work very well together. This selection becomes increasingly important if a user wants to equip multiple rooms throughout the home.

Voice assistants emerged from two distinct usage backgrounds. The first was solely as a controller for a smart speaker, while the second was a voice controller for a smart phone. Coming from the former category, Amazon’s Alexa was originally developed for the Echo smart speaker and then moved into the company’s FireTV devices. After Amazon opened up Alexa to third party developers, smart home proponents rapidly expanded its capabilities into features like controlling lights and thermostats by voice.

Google Assistant and Apple's Siri started out as the voice dialing features on Android phones and iPhones. The two systems let the user take calls, make calls, play music, and answer texts on the phone. Both systems are now focused on making their voice assistants more “smart speaker-friendly” by expanding their capabilities beyond the handset and into the home.

Other leading tech companies are also developing and deploying voice assistants for smart speakers, with Microsoft’s Cortana currently used in Harmon Kardon’s smart speaker product. Samsung’s Bixby voice assistant is reportedly in trials with at least one smart speaker vendor, although for now, it is currently only available on certain Galaxy phones.

Most smart home and voice assistant developers tend to segment the utility of voice assistants into three categories: top of the line; decent, but needs work; and still under construction. Both Alexa and Google’s Assistant are widely considered “top of the line” as the best voice assistants on the market, with Apple’s Siri and Microsoft’s Cortana considered decent, but needing improvement. Other voice assistants, such as Bixby and some being deployed by Chinese smart speaker vendors, are widely viewed as “still under construction.”

Market drivers
With demand spiking for smart speakers over the past year, a common question asked by those without a smart speaker goes something like “why do I need one of these products?” Or more specifically, what is the value proposition of the smart speaker? There are three broad value props for the smart speaker market:
• Usefulness/utility of smart speakers. The basic value of the smart speaker is in its utility, or more specifically its ability to perform a number of core functions (playing music, responding to voice commands, basic online search, controlling other devices in the smart home, etc.) extremely well.
• Directly tied to this utility is the price of smart speakers, with most products selling at retail for under $100. At this price point, the purchase of a smart speaker can be viewed as a relatively low risk purchase. In other words, if the user does not like it or find it useful, buying one is not going to break the household budget.
• While most consumers only use a smart speaker for a few simple functions, its ability to expand its reach – and its ability to control other devices throughout the home – is also a key driver for use and adoption. The intuitive use of voice control and its perceived “cool” factor (a sentiment highlighted by many smart home advocates) is also driving adoption.

• As homes become increasingly connected and more intelligent, most smart speaker vendors are counting on their products to be the centerpiece of the smart home. While not all smart home proponents want to see the smart speaker as the center of the ecosystem (some notable pay TV service providers are instead positioning the set-top box as the smart home centerpiece), many key players in the technology and telecommunications industries believe the smart speaker is a key steppingstone toward smart home development and adoption.

Market Challenges
The flip side to the growth of the smart speaker market are some key challenges that still need to be more thoroughly addressed by smart speaker vendors. Foremost among these challenges is security.
• Security concerns. Technology and business media publications have had a field day recently with stories of Amazon Echo products “spying” on users. One specific story highlighted how a family in Oregon found out Alexa recorded at least one private conversation and sent it to a contact in their address book.
Amazon confirmed that the incident did occur, and noted that the family’s Echo device woke up due to a word in background conversation sounding like "Alexa." The subsequent private conversation was heard as a "send message" request. When Alexa responded out loud "To whom?" the background conversation was interpreted as a name in the family’s contact list and a recording was sent to that person.
While this event seems to have been quite unusual, the fact that it did occur highlights the security and privacy challenges facing the smart speaker market. While smart speaker vendors are rapidly applying resources to better address security concerns, stories like this one are an obstacle to market growth.
• Another challenge to the market is a general lack of compatibility between smart speaker products and other connected devices. Leading vendors such as Amazon, Google and Apple have designed their products to interact and function with other compatible devices that often use proprietary software of operating systems. While Amazon has been aggressive in allowing third-party developers to integrate Alexa functionality into their products, for the most part Google relies on Android-based products to connect to its Home smart speaker, while Apple relies on IoS-based products for its connected ecosystem.
• Interconnection between the three ecosystems is tricky at best, and often just not possible. As Kagan does not see any serious drive for standardization among smart speaker products or technologies, this issue is going to remain a significant challenge for the near future.
• Related to compatibility with other products is the challenge of ecosystem complexity. While a key driver for the smart speaker market is the growing interest in the potential of the smart home, the caveat to that driver is how to best connect smart speakers with other devices in the home. Again, within the same smart speaker ecosystem (Amazon, Google, Apple, etc.), it is achievable. However, when third-party devices and services provided by cable operators, telcos or even security firms are added to the mix, the complexity of integrating a smart speaker into a broader ecosystem becomes a serious challenge.
• While Kagan identifies ecosystem complexity as a market challenge, we also recognize that some of the leading smart speaker vendors are purposefully designing their ecosystems to work best with their own proprietary platforms and products. This “silo” effect for a given ecosystem can offer the smart speaker vendor advantages over a more standardized and compatible ecosystem. As the smart home continues to develop, and as new revenue opportunities emerge from this development, having a proprietary ecosystem could turn out to be savvy business move on the part of some vendors.

Vendor Market Shares
Amazon, with its Echo smart speaker product line, dominated vendor market shares in 2017. On a global basis, Amazon accounted for just over 60% of smart speaker shipments, with Google’s Home smart speaker accounting for 31% of shipments.

Both Amazon and Google experienced strong growth in unit shipments in 2017. Amazon’s shipments grew from an estimated 5.5 million smart speaker products in 2016, while Google increased from an estimated 500,000 units shipped.

Alibaba and Xiaomi collectively shipped an estimated 1.5 million smart speakers last year, with almost all of their shipments occurring in China, while Sonos shipped an estimated 4000,000 units.

Apple’s HomePod smart speaker does not appear in these rankings, as it did not begin shipping until February 2018.

Worldwide smart speaker forecast
Since their introduction in 2015, demand for smart speaker products has risen impressively, making the product category one of the notable recent successes in the global consumer electronics market. After exceeding 31 million units in 2017, smart speaker shipments are projected to hit 50 million units in 2018.

Kagan’s longer-term forecast is based on several assumptions, to include:
• the increasing demand for smart speaker products
• their increasing utility, especially in regard to voice-enabled search and response capabilities
• the growing demand for smart home functionality
• an increase in demand for smart speakers in regions like Europe and Asia
This final point is particularly important, as current demand for smart speakers relies overwhelmingly on the North American market. In 2017, 83% of global smart speaker shipments were to North America.
Unit shipments are forecasted to surpass 70 million in 2019 and rise to 142 million by year-end 2022.

While a few countries in regions outside of North America are experiencing some solid demand for smart speakers, such as the United Kingdom, Germany, France, China, and South Korea, the key market to date has been the U.S. With the expanding availability of smart speakers in Asian and European countries, coupled with voice assistants integrating new language capabilities, we expect smart speakers to appear in an increasing number of households outside of the U.S.

Revenues for smart speaker products reached $2.52 billion in 2017 and are on track to exceed $4 billion in 2018. Average selling prices, or ASPs, for smart speaker products are generally below $100, with occasional sales promotions offering full-featured products for less than $50.
While most shipments have been concentrated among the sub-$100 price point, there are some notable higher-priced products available. These include the Sonos One, priced at $199 and Apple’s HomePod, priced at $350.
On a regional basis, North America produces the lion’s share of smart speaker revenues. In 2017, North American consumers accounted for an estimated $2.1 billion of the total $2.5 billion in product revenues.

Google And Amazon At The Forefront Of U.S. Renewable Energy Investment

Learn More

Discovering Scripps’ Programming Plans

Learn More

Banking & Financial Services
C&I Loan Growth Pops In Q2, But Tax Reform’s Role Remains Unclear

Jul. 31 2018 — Business loan growth popped in the second quarter, but bankers are hesitant to attribute the jump to tax reform or a broader turnaround in business spending.

The year-over-year increase in commercial-and-industrial loans increased to more than 5% for all banks in June, the highest figure in more than a year, according to Federal Reserve data. Smaller U.S. banks — defined by the Fed as those outside the 25 largest banks — posted double-digit growth for all three months of the second quarter.

Those numbers were artificially inflated by banks' acquisition of $24.9 billion of C&I loans from nonbanks. Accounting for those one-time acquisitions, organic C&I loan growth for smaller banks was still robust at 7% in June.

Ever since Republicans passed tax reform at the end of 2017, business optimism has been high and bankers have been hopeful the sentiment will trigger a rebound in business loan growth. C&I loan growth was less than 1% when tax reform passed.

Though C&I loan growth enjoyed a significant bounce in the second quarter, several bankers were not declaring victory. Numerous bank executives attributed the jump to an increase in merger-and-acquisition activity, not increased business spending.

M&T Bank Corp. said M&A activity was hurting its average loan growth, which declined by less than 1% on a quarter-over-quarter basis. The bank's CFO said businesses are selling significant assets and using the proceeds to pay down their loans.

One bank did say tax reform was boosting loan growth. SunTrust Banks Inc. reported an increase in the second quarter for its average performing loans figure, a turnaround from the first quarter when the figure declined on a linked-quarter basis.

"I think we are starting to see some of that [benefit from tax stimulus]," said Chairman and CEO William Rogers Jr. in the bank's earnings call.

But Rogers appeared to be in the minority. Several bankers said it was too early to tell whether tax reform was playing much of a role in the C&I loan growth. JPMorgan Chase & Co. reported a 3% quarter-over-quarter increase in its C&I loans in the second quarter and attributed the gain to M&A financing, not tax reform.

"We've yet to see the full effect of tax reform flow through into profitability and free cash flow," Lake said during the bank's earnings call.

Some bankers, including JPMorgan CEO Jamie Dimon, pointed to brewing trade wars as potential headwinds to loan growth.

Tariffs and trade-related issues are "probably the primary concern that we're hearing from customers right now," said Comerica Inc. President Curt Farmer.

Jeff Rulis, an analyst with D.A. Davidson, said he was not even sure the second-quarter C&I loan growth figures represented a notable change.

"I'm not convinced we're seeing a turnaround or significant pick-up. You have to take into account seasonal pick-up, and the first calendar quarter is generally slow," he said.

There is an argument that tax reform might actually be dampening loan growth. Rulis attributed high payoffs to the mixed results across the sector with some banks reporting robust loan growth by taking market share, contributing to others' more marginal results. Businesses are having an easier time making those payoffs thanks to tax reform, which freed up capital to pay down debt.

"One of the disadvantages of tax reform is you've both lowered the corporate tax rate and repatriated assets to the U.S. That's given more liquidity to the borrowers," said Peter Winter, an analyst with Wedbush Securities.

Year-over-year increases for total loans were up modestly, as weak commercial real estate loan growth moderated the gains from C&I. The 25 largest banks, in particular, reported soft commercial real estate loan growth with year-over-year declines in March, April and May — the first such drops since 2013. Several banks reported an intentional pullback from the sector due to credit quality concerns. Some pointed to nonbank competition as being particularly aggressive on both pricing and deal structure.

"I think banks, for the most part, are showing more credit discipline coming out of the financial crisis," Winter said. "Quite honestly, we're nine years into this recovery, so I think that's a prudent thing to do."

Learn more about Market Intelligence
Request Demo

Listen: StreetTalk Episode 27: Looking For The Cream Of The Crop In Bank Stocks

Jul. 30 2018 — Joe Fenech, head of equity research at Hovde Group, discussed current bank stock valuations, the growing importance of deposits in valuing franchises and the market's increased skepticism toward M&A, including transactions that appear favorable for the buyer.

No content (including ratings, credit-related analyses and data, valuations, model, software or other application or output therefrom) or any part thereof (Content) may be modified, reverse engineered, reproduced or distributed in any form by any means, or stored in a database or retrieval system, without the prior written permission of Standard & Poor's Financial Services LLC or its affiliates (collectively, S&P).

Banking & Financial Services
Loans And Deposits Continue Uphill Climb At US Banks In June

Jul. 26 2018 — Average total loans and leases at U.S. commercial banks increased by $44.10 billion to $9.347 trillion in June, according to the Federal Reserve's July 13 H.8 report.

Loan growth was driven primarily by a $19.8 billion increase in commercial and industrial, a $9.4 billion jump in real estate and an $8.3 billion increase in commercial real estate.

Average loans and leases at large commercial banks increased $18.7 billion month over month, while average loans and leases at small commercial banks were up $21.7 billion. Loans and leases at foreign-related institutions increased by $3.4 billion.

Meanwhile, average total deposits at U.S. commercial banks increased by $56.4 billion in June, compared to a $35.4 billion increase in May. Total deposits were up $448.4 billion from June 2017.

CECL Could Create Large Capital Shortfall For Community Banks

Learn More

Listen: StreetTalk Episode 24: "De Novos" Are Back. So Get Ready For More.

Learn More

Credit Analysis
Peeking Into The Future Without Staring At A Crystal Ball: Brexit Scenarios And Their Impact On Energy Firms’ Credit Risk

Jul. 24 2018 — After so many years of living and working in London, two years ago I applied for, and was finally granted, British citizenship. Imagine my surprise when, a few weeks later, the UK European Union referendum took place and the majority of voters opted for Brexit!

As a dual national, both European and British, I feel twice the pain of an uncertain future and sometimes I wish I had a crystal ball.

While it is hard to predict how the whole separation process will pan out, S&P Global Market Intelligence offers a new statistical model that allows users to understand how firms’ credit risk on either side of the ocean may change under multiple exit scenarios. The Credit Analytics Macro-scenario model covers the United States, Canada and European Union countries plus the United Kingdom (EU27+1). In addition, the model can be run via the S&P Global Ratings’ Economists macro-economic multi-year forecasts, tailored for this specific model and updated on a quarterly basis.1

Figure 1 shows the Economists’ forecasts of the inputs used in our statistical model for EU27+1, for year-end 2018, 2019 and 2020.

Source: S&P Global Market Intelligence (as of June 2018). For illustrative purposes only. L/S ECB Interest rate spread is the spread between long-term and short-term ECB interest rates. Y-Axis is % of; GDP Growth, Stoxx50 Growth, Interest Rate Spread or FTSE100 Growth, depending on the correlating symbol as described in the key.

The expectation is for economic growth to slow-down in the EU27+1. This will be accompanied by progressive monetary policy tightening and a volatile performance of the stock market index growth. This view is aligned with the baseline scenario included in the European Banking Authority (EBA) and the Bank of England (BoE) 2018 stress testing exercise that “[…] reflects the average of a range of possible outcomes from the UK’s trading relationship with the EU”.2

Figure 2 shows the evolution of the median credit score of Energy sector (left panel) and Utility sector (right panel) large-revenue companies in EU27+1, obtained by running the economists’ forecasts via the Macro-scenario model.3 The median score for 2017 is generated via S&P Global Market Intelligence’s CreditModelTM 2.6 Corporates, a statistical model that uses company financials and is trained on credit ratings from S&P Global Ratings.4 The model offers an automated solution to assess the credit risk of numerous counterparties, globally. The scores are mapped to a numerical scale where, for instance, bb- (left panel, left scale) is mapped to 13.0; a deterioration by 1 notch corresponds to an increase of one integer on the numerical scales.

Figure 2: Evolution of the median credit score of Energy and Utility sector companies in EU27+1, based on S&P Global Economists’ macro-economic forecasts run via the Macro-scenario model.

Source: S&P Global Market Intelligence (as of June 2018). For illustrative purposes only.

Starting from 2017, we see a higher level of credit risk in the UK (red line) than in EU27 (blue line); in subsequent years, the median credit risk increases on both sides of the Channel but the “risk fork” between the UK and EU27 tends to widen up at the expenses of the UK, for both sectors.5

Despite the fact that the median credit score may not change sizably between 2017 and 2020, remaining below half a notch overall in all cases, it is worth keeping in mind that the probability of default (PD) associated with a credit score changes in line with the economic cycle, and thus increases (decreases) during periods of contraction (expansion).

In our model, we account for this effect by first mapping the credit score output to a long-run average PD; next we scale it via a “Credit Cycle Adjustment” (CCA) that looks at the ratio between the previous year and the long-run average default rate historically experienced in S&P Global Ratings’ rated universe.6 If we adjust the long-run average PD via the CCA, we can easily identify potential build-up of default risk pockets in different countries within the EU27+1 as time evolves, as shown in the animations within Figure 3. Green refers to a lower PD than 2017, orange refers to a higher PD than 2017, and red refers to a PD breaching a pre-defined threshold (4.5% for Energy Sector and 0.3% for Utilities sector).7

Figure 3: Potential pockets of default risk in Energy and Utility sector companies in EU27+1, based on S&P Global Economists’ forecast.

Energy Sector Utility Sector
Default Risk in Energy map Default Risk in Utilities map

Source: S&P Global Market Intelligence (as of June 2018). For illustrative purposes only. Green refers to a lower PD than 2017, orange refers to a higher PD than 2017, and red refers to a PD breaching a pre-defined threshold (4.5% for Energy Sector and 0.3% for Utilities sector).

With the Macro-scenario model, we aimed for a user friendly model, and took into account the strong economic ties within EU27+1, the existence of a common market and the circulation of a shared currency in the majority of the EU countries, in order to select a parsimonious yet statistically significant set of inputs (just imagine otherwise forecasting multiple macro-economic scenarios for 28 individual countries, over multiple years).8

Readers may wonder how the model differentiates the evolution of credit risk by country if it uses a limited set of aggregate macro-economic factors (e.g. EU28 GDP growth, etc.) across EU27+1. Nine separate sub-models were actually optimized, based on economic commonalities and historical evolution of the S&P Global Ratings transitions in those countries, to account for the existence of different EU “sub-regional” economies (for instance Nordic countries as opposed to Eastern European countries). For the UK, we went one step further, by explicitly including a market indicator, the FTSE100, as a precautionary measure given a potential “full decoupling” of EU27 and UK economies in the near future.

Well, so far so good, at least in the case of a “soft” Brexit! But what if we end up with a “hard” Brexit?

The EBA and the BoE 2018 stress testing exercise include a stressed scenario that “[…] encompasses a wide range of economic risks that could be associated with {hard} Brexit”.9 The scenario corresponds to a prolonged recessionary period, with negative GDP growth for several years and a generalized collapse of the stock markets, similar to what happened during the 2008 global recession. Unsurprisingly, the median credit score output by our macro-scenario model companies significantly deteriorates for both Energy and Utility sector. Figure 4 shows the build-up of potential default risk pockets and their evolution over time, under stressed economic conditions, depicting a bleak view over the length of time needed for a recovery of these sectors.10

Figure 4: Potential pockets of default risk in Energy and Utility sector companies in EU27+1, based on EBA’s and BoE’s 2018 stressed scenario.

Energy Sector Utility Sector
Default Risk in Energy map Default Risk in Utilities map

Source: S&P Global Market Intelligence (as of June 2018). For illustrative purposes only. Green refers to a lower PD than 2017, orange refers to a higher PD than 2017, and red refers to a PD breaching a pre-defined threshold (4.5% for Energy Sector and 0.3% for Utilities sector).

I do not have yet a crystal ball to predict the future, e.g. whether petrol will cost more or less, or whether I will be paying higher utility bills in the UK as opposed to (the rest of) the European Union, but S&P Global Market Intelligence’s Macro-Scenario allows gauging potential credit risk changes in individual countries, under a soft or a hard Brexit scenario. More in general, the Macro-Scenario model offers a quick, scalable and automated way to assess credit risk transitions under multiple scenarios, thus equipping risk managers at financial and non-financial corporations with a tool that enables them to make decisions with conviction.


1 The macro-economic forecasts will become available on the S&P Capital IQ platform from 2018Q4. S&P Global Ratings does not contribute to or participate in the creation of credit scores generated by S&P Global Market Intelligence.

2 Source: “Stress Testing Exercise 2018” available at The baseline scenario is the consensus estimate among EU27+1 Central Banks.

3 The results of this analysis depend on the portfolio composition. In addition, other industry sectors may react differently from the Energy and Utility sectors.

4 S&P Global Ratings does not contribute to or participate in the creation of credit scores generated by S&P Global Market Intelligence. Lowercase nomenclature is used to differentiate S&P Global Market Intelligence credit scores from the credit ratings issued by S&P Global Ratings.

5 The 2017 median score for the Utility sector is better than the score for the Energy sector, due to the inherently higher risk of companies in the latter.

6 An optional market-view adjustment is available within the macro-scenario model. In our analysis, we did not include this adjustment, for the sake of simplicity.

7 4.5% (0.3%) is close to the historical long-run average default rate of companies rated B- (BBB-) by S&P Global Ratings.

8 This is also one of the reasons we found it unnecessary to include oil price for the modelling of credit risk of the energy sector in EU27+1, as we found the stock market growth was sufficient.

9 Source: “Stress Testing Exercise 2018” available at The baseline scenario is the consensus estimate among EU27+1 Central Banks. Curly brackets refer to the author’s addition.

10 We adopt the same colour conventions as in Figure 3.

Understanding Drivers Of Credit Risk

Learn More

Macro-Scenario Model: Conditioning Credit Risk Transitions On Macro-Economic Scenarios

Learn More