Last week’s tech stock rout saw Alphabet’s record single-day market cap loss of $500Bn+, 12%. The shockwaves cascaded across the tech sector, sending the Nasdaq tumbling to a nine month low and dragging down the sector. Beneath the financial headlines was some insightful Tech Analyst views of the market and product dynamics. Here are my reflections from these over the weekend.
When Alphabet released its financial update, market observers were initially greeted by surface level strength: soaring cloud computing revenues +103% year-on-year, and showcasing strong adoption of its Gemini AI models. However, lurking beneath these headlines was a reality that sent tremours: AI search summaries (AI Overviews) were eating into Google’s core revenue of search advertising. Within hours, we had a possible landmark turning point in investor sentiment regarding AI, just weeks after the frenzy around the Anthropic IPO.
For decades, Google’s search engine operated as the most lucrative internet based business model. We typed queries, Google pinged back and displayed the blue links alongside targeted text ads, and advertisers paid when we clicked through. The introduction of AI-generated summaries at the top of search results has fundamentally disrupted this, as Gemini provides direct, comprehensive answers to users’ search questions. Users no longer need to click on underlying website links or sponsored advertising slots.
This shift presents Google with their own version of the Innovator’s Dilemma: to remain competitive against AI-native search alternatives like Perplexity, Google is forced to cannibalise its own primary profit stream. What rattled investors was its acknowledgment that the increasing use of AI Mode in Google search which produces chatbot-like answers, was yielding lower monetisation than traditional search.
Contagion tipped the tech-heavy Nasdaq index, and a broader sell-off took hold. Investors began taking a harder, critical look at the massive CapEx being poured into AI infrastructure versus the tangible cash returns generated.
Microsoft had reported Q3 just days after announcing plans to cut its workforce by 20% (60,000) as it pours billions into AI. Reaction to their update was more muted, with shares falling less than 1%. CEO Satya Nadella said the use of Copilot, Microsoft’s AI chatbot and tool, had seen automated tasks nearly double since July as daily users grew 500% year-on-year.
Similarly, Amazon saw its shares drop 6% following earnings that missed investor expectations. Like Alphabet and Microsoft, Amazon has committed billions for AI datacentres and server infrastructure. Alongside quarterly sales growth of 11%, down from 26% in the previous quarter, Amazon shared a staggering cash flow figure of negative $42bn: Amazon is spending more money than it makes, and that spending has increased six-fold over three months. Amazon suggested it saw an opportunity making and selling its own chips, and from its AI developer platform Bedrock, offering paying users access to AI models and tools.
Does this broad market anxiety reflect a fundamental shift in investor psychology towards AI? Recently we’ve seen investors ignoring the fundamentals of economics and mathematics, rewarding tech companies simply for announcing increased AI spending and data centre expansions. Has investor patience evaporated, replaced by demand for near-term monetisation, a route to profitability and cash runway protection?
In its analysis of the market rout, The Economist frames Google’s $500Bn wipe out as a line in the sand. The threat Google search faces from AI is structural. Historically, Google enjoyed a virtual monopoly in search due to network effects, vast data moats, and deep integrations across web browsers and mobile operating systems, such default search deals on Apple’s iOS. The Economist highlights that Google is trapped between a rock and a hard place: if it slows down AI integration to preserve ad revenue, it risks losing search market share; if it accelerates AI integration, it accelerates the decay of its most profitable business.
Are we at a macroeconomic and infrastructure crossroads? The market sell-off underscores a growing tension: the massive divergence between consumer software monetisation and hardware/infrastructure costs. However, with interest rates remaining sensitive to inflationary pressure and cash flows shrinking under the weight of CapEx, investors are questioning whether Big Tech can generate enough revenue and profit to justify billions in annual capital deployment.
As all titans continue to invest gargantuan sums in data centres, the health of their traditional business model cash cows becomes all the more important. There is also investor concern about the data centre companies, known as neocloud firms. Many question whether neocloud firms are financially sound and what their recent stumbles tell us about the state of the AI economy.
Neocloud firms often finance themselves by signing long-term lease contracts with major tech customers who need access to computing power. Firms can then borrow money using those contracts – plus the AI chips they purchase – as collateral. Some jittery investors appear to be having second thoughts. For example, CoreWeave has lost a quarter of its value over the past month.
But markets don’t trade the quarter that was, they trade the quarter that will be. The four largest US cloud providers carry roughly $2.3Tn on contracted revenue backlog, with $1Tn in OpenAI and Anthropic. This means the largest capital cycle in the history of tech is underwritten by two private companies that do not make money.
Anthropic is making more revenue right now than ever before, but the problem isn’t now, it’s the future. Anthropic is committed to spending $800bn on data centre infrastructure over the next decade. OpenAI has a similarly-sized commitment. Critically, these are take-or-pay contracts, which means they have to pay whether they use the data centres’ compute capacity or not. Secondly, they haven’t started paying yet, because the data centres aren’t built! Anthropic’s valuation is down to them making $200Bn a year in revenue in two years: they’re currently making $40Bn.
In short, the maths currently don’t add up. Still with me? So much to absorb, consider, reflect and ask So What? It’s fascinating, and a little concerning at the same time. Here are my ten key takeaways for Manchester tech startup founders derived from last weeks’ Wall Street shenanigans:
1. Changing user experience and expectations: Search is evolving from page navigation – find me a website that sells shoes – to direct task completion – find the best running shoes for flat feet under £120 and summarise three top reviews. AI models synthesize information, eliminating the need for users to navigate away.
2. Compressing unit economics: Serving a standard Google search query costs a fraction of a cent in compute power. In contrast, running a complex LLM inference query requires GPUs, immense power, and specialised server cooling, making each AI response significantly more expensive to serve while generating less ad revenue.
3. Enterprise AI consumption is exploding. Google Cloud revenue doubled year-over-year, proving that B2B buyers have moved past AI pilots into full-scale production. Startup ventures building AI integrations or enterprise automation can capitalise on an active buyer market.
4. Aggressive infrastructure spend keeps cloud costs competitive. Google’s forecast for 2026 CapEx is $200Bn, focused on AI infrastructure and custom TPU compute. The intense capital competition among Big Tech keeps cloud credits, specialised hardware access, and hosting pricing competitive. There’s a deal to be had with smart negotiation between AWS, GCS and Azure.
5. Margin and cash concerns signals capital efficiency is back in demand. Despite 25% top-line growth, Alphabet’s net profit dropped 25% ($26.2Bn) due to heavy AI infrastructure spend. Investors are wanting clear paths to profitability and capital efficiency from early-stage ventures. High CapEx meets investor impatience: Wall Street has pivoted from rewarding aggressive AI investments to demanding visibility on revenue returns and margin expansion.
6. TPU diversification unlocks options beyond Nvidia. Alphabet reported that TPU sales contributed significantly to cloud growth. Deep Tech and ML ventures now have viable alternative hardware pipelines to train and run inference on, easing dependence on scarce GPUs.
7. API Token usage is booming. Google reported first-party model APIs processing over 32Bn tokens per minute. Developer first startups building middleware, observability tools, or prompt orchestration frameworks have a massive TAM as API consumption surges.
8. AI is triggering the Innovator’s Dilemma in full effect: Incumbents like Google are forced to aggressively deploy AI features that destroy their existing, highly profitable business models to prevent losing market share to competitors. Take a moment to look at this potential impact on your own value proposition and business model. Is this an opportunity or a threat?
9. ‘Other bets’ losses highlight core business focus. Alphabet’s ‘Other Bets’ division posted an operating loss of $2Bn on $372M revenue. This serves as a warning for founders against taking moonshot projects to market without strong unit economics or a proven core revenue stream.
10. Talent opportunities as Big Tech tighten cost focus. Big Tech’s pivot to agent first development internally means a sharp focus on headcount. Scale-ups can leverage this market dynamic to leverage senior engineering and ML talent availability with competitive share options and salaries.
Summary. The sell-off erased roughly $2Tn from the Nasdaq-100. The primary trigger was mounting investor anxiety over massive, front-loaded AI infrastructure spending alongside disappointing earnings figures. Tech giants like Alphabet and Amazon raised alarms with surging CapEx and negative free cash flow. Broader hardware, software, and semiconductor names also fell, as Wall Street sharply reevaluated immediate returns on AI investments.
Do the numbers add up? Is this a pause and a correction, or a red flag for AI? Either way, I see the road ahead as a continued opportunity for Manchester tech ventures that focus on the fundamentals of PMF, talent acquisition and diligently navigating their cash runway along the route to profitability.