Talking Tech

Talking Tech is produced by the Progressive Tech Team of George O’Connor, Ian Robertson and Gareth Evans. Our aim is to bring you up to date with the tech news cycle each week. We comment via blog and podcast on the slings and arrows of the sector at a time of huge change.

<< Back to Talking Tech archive

March 20, 2024

Politics is eating software: EU AI Act trumps the tech economy

Apologies Marc Andreessen: Software is being gorged on by politicians intent on devouring it’s inexorable rise. The gatekeepers are no longer Alphabet, Amazon, Apple, ByteDance, Meta and Microsoft (apologies EU) but the political class. Tech eyes are focused on Nvidia’s industry jamboree and SXSW but the big decisions are happening elsewhere. To wit: the EU AI Act puts guardrails on the development of AI, America forces ByteDance to divest TikTok, interest rate decisions are pending in the US (20/3) and UK (21/3). As the UK 10-year flips and flops around 4% the now expected ‘higher for longer’ (cost of capital) translates to ‘lower for longer’ (valuation) and so tech woes will persist. Note: UK tech salary rates are crumbling (see inside), job losses mount (see inside) and undermine the ‘higher for longer’ rationale. But, our financial round-up illustrates continued dogged company performance. This week we have positive updates from Bytes, Computacenter, Eagle Eye, Team Internet, Seeing Machines, SThree and Trustpilot and we scan the outlook statements for sentiment (below). Of our UK universe only Focusrite posted poor news. All in, Financial results continue to nibble away at the doubters. Learn more inside.

The new EU AI Act

The new EU AI Act creates a risk-based classification for AI systems. AI providers have obligations depending on the level of risk their AI system generates. Those with an unacceptable level of risk are prohibited by year end.

The Act comes in the wake of the Digital Markets ActGDPR and the WEEE Directive. In our view, these were positive in shaping the development of the tech industry, but we confess to having reservations about the AI Act. Why donate market to US tech in their drive for AI dominance and disruption? LLM suppliers are not going to develop different models for different regions so the EU legislation has global implications. Compare with the UK approach. In November, at the global AI Safety Summit held in London, there AI developers made a (non-binding) commitment to work with governments to test new models before they were released to help manage the risks of the technology before it reaches the public. Note: The UK is home to a third of Europe’s AI start-up companies and twice as many as any other European country (source: great.gov.uk). The UK has more than 1,300 AI companies with a collective turnover of c £1.47bn (source: great.gov.uk).

Flag waving interlude: Congrats to brit Mustafa Suleyman, DeepMind co-founder, hired (19/3) by Microsoft to run its new consumer AI unit. FT has the story

The Risk-based Classification

The EU AI Act introduces a risk-based classification scheme for AI applications. The main criterion is the level of risk posed by the AI application to individuals or society as a whole. The classification ranges from minimal risk to applications which are banned entirely.

  • Unacceptable Risk: Some AI applications such as social scoring systems or manipulative systems potentially leading to harm are outlawed completely.
  • High Risk: High-risk applications include services directly affecting citizens’ lives (e.g., evaluating creditworthiness or educational opportunities, applications applied to critical infrastructure). They will have to be put through strict assessment regimes before they can be put on the market. Businesses need to consider whether their existing or planned AI application might be considered “high risk”. The EU will update and expand this list on a regular basis.
  • Limited Risk: Other AI applications still carry obligations with them, such as disclosing that a user interacted with an AI system. Best practices related to data quality and fairness are essential even in this risk regime. Some examples are image and video processing, recommender systems, and chatbots.
  • Minimal Risk: Applications such as spam filtering or video games are deemed to carry a minimal risk and as such they are not subject to further regulatory requirements.

Points of note

The EU uses a wide AI definition

The EU definition of “artificial intelligence” attempts to be future-proof and so it covers a wide range of data analysis. Our concern is that the definition based on OECD formula is so broad that many technologies used today fall under its regulations. According to the EU; “Artificial Intelligence system’ means a machine-based system that is designed to operate with varying levels of autonomy and that can, for explicit or implicit objectives, generate outputs such as predictions, recommendations, or decisions that influence physical or virtual environments”. So, on the banned list is social scoring with “techniques that manipulate people in a way that impairs their autonomy, decision-making and free choices”. Examples of Social Scoring in use today include:

  • Insurance companies use it to allow operators to gather information about people’s behaviour. For example the car insurance industry calculates fees charged based on information received from a tracking device deliberately installed in a car.
  • Restaurateurs create blacklists of people not allowed to enter a bar or a restaurant due to previous mis-behaviour.
  • Sharing economy services, (think Airbnb, taxi, delivery services) have some kind of scoring system that evaluates customers from various points of view.
  • The three main UK credit reference agencies (Experian, Equifax and TransUnion) – each hold different personal and financial details about individuals which forms their credit history and it used to create a credit score. Each has its own method of calculating the score. You could even have a ‘good’ score with one and ‘fair’ with another.

EU AI Office

This is a new initiative to coordinate compliance, implementation, and enforcement of the AI Act. EU citizens can submit complaints about AI systems when they suspect they have been harmed by AI systems, and can receive explanations on why the AI systems made decisions they did. We are reminded of the ‘Right to be Forgotten’ from GDPR. Positively, this gives people some agency, negatively it requires a high level of AI literacy.

Non-compliance

The AI Act imposes significant penalties for non-compliance with the prohibited systems provisions, fines up to €35m, or 7% of global turnover.

A challenge of legislators is the pace of change

Both the pace of change in GenAI technology and the pace of adoption has been awesome. For example, at GTC (18/3) Nvidia unveiled the Blackwell GPU architecture this is the first development in chip design since the arrival of the Hopper architecture, only two years ago. Blackwell offers out-sized improvements over its predecessor the 20 PetaFLOPS of AI performance that Blackwell offers is 4x faster on AI Training workloads, 30x faster on AI Inferencing workloads and, most notably, up to 25x more power efficient than its predecessor. Attempting to ring-fence a technology which is evolving at such a rapid pace can only be challenging.

Meanwhile there are persistent cracks in the tech macro

The Bank of England is likely to keep rates unchanged and talk about the need to see more data before committing to rate cuts at Thursdays (21/3) MPC meeting. The narrative is likely to be about continued elevated wages and services inflation, echoing comments likely made by the Fed later today (20/3). However, the data says that the previous tight tech labour market and tech wage growth have both unwound. This is reflected in SThree Q1/24 Trading Update where net fees for Contractors was -2% Y/Y and for Permanent staff was -21% Y/Y. For the tech sector at least rate cuts cannot come fast enough. Note:

  • Wages. Our latest data run shows the largest fall in UK IT salary rates (see below). Full time rates in March £56,939, -8.1% month-on-month sequentially, with contractor rates, £531.2, -4.7% sequentially. This makes the YTD average down -7.6% Y/Y and -3.1% Y/Y for full time and contractors respectively.
  • Job reductions continue. Data from the Layoffs.com once again shows the continued deterioration in the tech economy. YTD2024 saw 209 tech companies layoff 50,312 employees (see below).
  • Hiring intentions weaken. The latest Manpower Employment survey concludes that hiring intentions are falling. March data (12/3) showed the 22% Global Net Employment Outlook at 22% which was -2% Y/Y and -4% vs. Q1 2024.
  • Employee sentiment. Our latest Glassdoor scan also shows the Happy index dip, being 3.79 vs 3.80 sequentially (see below).

Can we blame AI for everything – why not?

More than one-third (37%) of business leaders say AI replaced workers in 2023, according to a report from ResumeBuilder, but the impact of AI on employment is only a known-unknown. AI is cited as an influence on job reductions, but the economy is the main named reason (see below). In truth much of the blame is the over-hiring in the wake of the pandemic as companies panicked but this will gather pace. The IMF argues that c60% of jobs in advanced economies are exposed to AI of which half may be negatively affected as AI’s ability to affect highly skilled jobs means that advanced economies face greater risks from the technology.

Financial news update

Reviewing the latest sector outlook statements – Six out of Seven are positive

Bytes (Mkt Cap £1.29bn/PE 30.4x/Share 1 year 45.8%)

Full Year Trading Update (18/3): “continuing ability to deliver sustained double-digit growth”

“The Group once again delivered growth comfortably in double digits in its key metrics of Gross Profit and Adjusted Operating Profit, and cash conversion in line with the Group’s target of 100%, resulting in a cash position of approximately £89m at the year end. . . This reflects the very strong demand for software and IT Services we continued to see from both corporate and public sector clients, despite the well-documented macroeconomic headwinds. BTG remains confident in its continuing ability to deliver sustained double-digit growth.”

Note: Bytes also updated on the investigation of its former CEO Neil Murphy following his resignation (21/2/2024) of BTG’s former CEO, Neil Murphy.

Computacenter (Mkt Cap £3.61bn/PE 17.5x/Share 1 year +46.61%)

Final Results (20/3): “we expect 2024 to be another year of progress with growth weighted to the second half”

PLC Award CEO of the Year Mike Norris: “We delivered our nineteenth consecutive year of growth in adjusted earnings per share, outperforming our markets in 2023, as our large customers continued to invest heavily in new technology. We managed an uncertain macroeconomic backdrop and inflationary pressures effectively, reduced our inventory significantly, resulting in a record net cash position. Overall we expect 2024 to be another year of progress with growth weighted to the second half, while continuing to invest for future growth.”  

Note: (i) Numbers include a 15 day swing in inventory days to 13 i.e. the lowest since pre-Covid (ie 2019 10 inventory days)  (ii)) CCC comments that “Given the strength of our balance sheet we continue to evaluate a number of capital allocation options”. (iii) This is the nineteenth consecutive year of adjusted EPS growth – roll on twenty years.

Eagle Eye (Mkt Cap £170m/PE 144x/Share 1 year 1.3%)

Interim Results (19/3): “In line with expectations”

“We have entered the second half of the financial year with good momentum. Our sales pipeline has increased considerably versus this time last . . . Trading since the Period end has continued well, providing confidence in the delivery of another year of profitable growth, in line with the Board’s expectations.”

Focusrite (Mkt Cap £169m/PE 9.4x/Share 1 year -55%)

Trading Update (18/3): “challenging market conditions”

A negative trading update (FY ending 31 August 2024) saw the “challenging market conditions highlighted in January’s AGM statement continued throughout February and into March”. FY24 revenue is now anticipated to be “not less than £155m”, EBITDA in the range £27 – £30m with a heavier weighting towards H2 than in FY23.  Revenues for HY24 are anticipated to be not less than £75m, from £86.2m Y/Y.”

Seeing Machines (Mkt Cap £210m/PE na/Share 1 year -15.9%)

Interim Results (18/3): “results to be in line with expectations”

“The Group remains well-placed to deliver continued progress in the year ahead, with a typical weighting to the second half, and the Board retains its expectations that financial performance for FY2024 will be in line with consensus.”

SThree (Mkt Cap £572m/PE 10.2x/Share 1 year -4.0%)

FY24 Q1 Trading Update (19/3): “Performance for FY24 currently expected to be in line with market expectations

“Once again, we have delivered a good performance against a strong comparative and within a market environment that remains difficult from a new business perspective. Whilst the sentiment we are reporting is much the same as the prior period, the strength of our Contract extensions continues to be a particular highlight, demonstrating our clients’ need to retain critical STEM skills and flexible talent.

Team Internet (Mkt Cap £351m/PE 19.3x/Share 1 year +3.8%)

Final Results (18/3): “to meet current market expectations”

“We confirm our guidance that the Group is confident in its ability to meet current market expectations. Our established brands, market leadership, cutting-edge technology, the assembled talent and robust cash flow will allow us to translate our current success into future success.”

Note: Team Internet posted an excellent slide on its AI initiatives (see below).

Trustpilot (Mkt Cap £850m/PE na/Share 1 year +123.9%)

Final Results (19/3): “confident in the ability to deliver sustainable growth and long-term margin improvement

“We delivered a strong performance in 2023, as we accelerated our move into profitability, delivering adjusted EBITDA ahead of expectations. The bookings growth we achieved in 2023 and the ongoing momentum in the business underpins our confidence in continuing to deliver mid-teens constant currency revenue growth, and we also expect to achieve further operating leverage in the current financial year. The Board is confident in the Company’s ability to deliver sustainable growth and long-term margin improvement, as we expand to capture the significant global opportunity ahead.”

The data

UK Tech Full time salary and contractor rates (£)

Source: ITjobswatch.co.uk, Analyst

Manpower Employment Outlook, since 2020

Source: Manpower Employment Survey

Reasons for US Job cuts in 2023

Source: Challenger, Gray and Christmas

The Happy Index: Glassdoor rating

Source: Glassdoor

Team Internet illustrates how AI is woven into the company.

Source: Company data

This communication is provided for information purposes only, and is not a solicitation or inducement to buy, sell, subscribe, or underwrite securities or units. Investors should seek advice from an Independent Financial Adviser or regulated stockbroker before making any investment decisions. Progressive Equity Research Ltd (“PERL”) does not make investment recommendations.

Opinions contained in this communication represent those of PERL and/or our affiliates at the time of publication and PERL does not undertake to provide updates to any opinions or views expressed. PERL does not hold any positions in the securities mentioned in this communication, however, PERL’s directors, officers, employees, contractors and affiliates may hold a position,  and/or may perform services or solicit business from, any of the companies or related securities mentioned.

Any prices quoted in our research are as at the previous day’s close.