Omdia: AI budgets to rise in 2027 as supply chains, sovereignty reshape tech
Omdia expects 59% of organisations to raise AI budgets by 10% or more in 2027, part of four forces the analyst house says will reshape the industry, it said on 30 September.

On 30 September, Omdia published four forces it expects to reshape technology in 2027: AI monetisation, persistent supply chain disruption, digital sovereignty and Physical AI. The Light Reading write-up of the press release states the shift plainly: the industry is moving from what technology can do to what it delivers, economically and operationally.
The headline number is 59%. That is the share of organisations Omdia says expect their AI budgets to increase by 10% or more in 2027. Pressure to show returns will intensify.
Enterprises will judge AI by ROI and productivity gains rather than technical superiority, according to Omdia. That is a different test from the model benchmarks that dominated the past three years. The question becomes which business models turn adoption into sustainable returns, across infrastructure, software, services and devices. Hardware delays already affect 60% of PC channel partners, Omdia says, which gives the supply chain argument a current number rather than a forecast.
Brussels writes the rules, Washington signs a pact
The European Commission's AI Act remains the reference point for binding rules, but the political weather has changed. On 29 September, EU technology chief Henna Virkkunen told the European Parliament the Commission is preparing guidance on high-risk AI systems, with a first draft expected by the end of the year, Euractiv reported. The same week, EUobserver reported that member states are being urged to go beyond the AI Act to protect people from abusive surveillance, a sign that the enforcement debate is not settled. Digital sovereignty is one of Omdia's four forces; more than 100 countries now pursue sovereignty initiatives, the analyst says, but complete independence will rarely be practical.
Across the Atlantic, the contrast is sharp. At a White House summit on 30 September, President Trump and technology executives signed a voluntary agreement to self-police AI development, according to NPR. The BBC's three takeaways from the event noted the pact is not legislation. Al Jazeera reported the same accord, describing the language as "robust" safeguards; the Telegraph reported that Trump claimed tech bosses agreed to drop calls for regulation, which the executives have not confirmed on the record.
"The question is increasingly shifting from whether organisations need greater digital sovereignty to how much control they actually need," Omdia wrote in its 30 September release.
That framing matters for EU policy. The bloc's AI Act, the General Data Protection Regulation and the Digital Markets Act all assume that legal control over data and models is achievable. Omdia's point is that the cost of full control rises with the scope. Ireland's data protection regulator said on 29 September that tech giants cannot use private messages to train AI, a decision that tests the boundary between platform data and model training. Google has appealed EU orders to open search data to AI rivals, according to ET Enterprise AI, and the EU's General Court will hear the case in January.
Physical AI moves the goalposts
Omdia's fourth force, Physical AI, may be the most consequential for regulators. Global humanoid robot shipments are expected to exceed 700,000 units by 2030, the analyst says, spanning vehicles, buildings, energy infrastructure and workplaces. Once intelligence leaves the screen, liability, safety and integration rules written for software become harder to apply. The AI Act's product-safety provisions will meet hardware that learns in the field.
Not everyone is convinced the returns are there. A report by Lighthouse Reports on data centre silence, published 30 September, argues that the environmental cost of AI infrastructure is being hidden in planning processes. Politico reported the same day that the EU is accused of hiding the environmental impact of data centers, while Pinsent Masons noted EU data centre rating plans that signal greater scrutiny of AI infrastructure sustainability. Those stories sit awkwardly with the investment numbers Omdia cites.
The Linux Foundation's Technical Advisory Board election, meanwhile, shows that not all governance is a matter of state regulation. LWN.net reported on 30 September that nominations close on 7 October, with five seats to fill, including the one vacated by the death of Dan Williams. It is a reminder that the technical community runs its own processes, often faster than legislatures.
On security, the browser is now the battleground. The Hacker News reported on 30 September that most breaches begin and end in a browser session, and that ClickFix, where users are tricked into pasting malicious commands, accounted for 47% of observed attacks in Microsoft's Digital Defense Report, and 52% of Push detections in Q2 2026. Roughly one in two phishing attacks is delivered outside email, and 89% of phishing domains live for fewer than two days. The AI Act's cybersecurity requirements will land on organisations that are already losing this fight.
The through-line is cost. Omdia says AI budgets are rising, but so are the bills for sovereignty, supply chain resilience and physical deployment. The EU wants rules; the US wants voluntary pledges. Whether either approach produces sustainable returns is the test Omdia has set for 2027.
Sources
3- 01Four forces set to reshape technology in 2027EN
- 02The Linux Foundation Technical Advisory Board 2026 election approachesEN
- 03Know Your Enemy: Browser-Based Attack Techniques in 2026EN
All figures and quotations in this text come from the sources listed below.
Content prepared by the editorial team with AI assistance.
Comments
0- No comments yet — be the first.