Insider · Tech Breaking
PwC's US and India shake-up is a warning to anyone still treating offshore delivery as safe from AI. The work is moving up the chain, or it is moving out. If you want to see what a professional services firm looks like when it decides to disrupt itself before someone else does, look at PwC.
The Financial Times reported that PwC's US and India firms are preparing a joint venture of about 40,000 people, combining PwC US's offshore acceleration centres with PwC India's consulting business. The agreement is scheduled to be signed on September 14, 2026, with completion expected in the first half of 2027, subject to regulatory approval. Sanjeev Krishan, chairperson of PwC in India, is expected to lead the venture.
That detail matters. This isn't a side project being parked in a back office. It puts a large Indian consulting and delivery workforce closer to PwC's US client machine, at a moment when the old offshore model is under pressure from AI tools that can do more of the repeatable work junior teams used to handle.
Here is the blunt version. PwC built a large part of its US delivery model around acceleration centres in India, where teams could help produce client work at lower cost than US staff. That worked when the job was hours of manual analysis and document production.
AI has started eating into exactly that layer. A cheaper worker is still more expensive than software if the task is narrow enough. That is the warning shot.
PwC is changing the unit before AI changes it for them The stated aim is to make the combined firm more useful to multinational clients by joining cheaper delivery capacity with AI, technology and engineering skills. Fine. But the real story is sharper than the usual language about integration.
PwC is trying to turn people who execute defined tasks into people who can build, deploy, monitor and fix the systems that automate those tasks. That is a hard turn. A junior consultant who once cleaned data or built decks now has to prove they can work around models and client systems.
If they can't, the economics get ugly fast. PwC has already been moving in this direction. In April, the Financial Times reported that PwC UK would merge its risk and consulting divisions as part of a wider push to standardise services across the network and use more shared staff in places such as India.
The enlarged UK consulting division was reported to have about 4,600 people and roughly £1.1 billion in annual revenue. Marco Amitrano, PwC UK's senior partner, said the decision was about global alignment. You can see the pattern.
The old Big Four network model, with national partnerships stitched together under a global brand, was built for a slower world. Clients now want one answer from the firm, not a tour through every internal service line before the work begins. The rest of consulting is already moving Look at Accenture.
The company said in June that it had agreed to buy a majority stake in Dragos and all of runZero and NetRise at a combined enterprise value of about $4.175 billion. That is not a random cybersecurity deal. Dragos protects operational technology used in power grids, factories, pipelines and data centres, exactly the physical infrastructure where AI-driven threats are becoming more serious.
Accenture has also been reshaping its workforce around AI. Its own filings put headcount at about 799,000 at the end of May 2026, up from about 779,000 at the end of August 2025, so the simple story is not mass shrinkage. The better reading is churn.
Some roles disappear, while others get built around AI, data, cyber and engineering. Deloitte is making the India side of that bet even more directly. People Matters reported in April that Deloitte plans to hire 50,000 professionals in India, citing South Asia chief operating officer Nitin Kini, with a focus on AI and emerging technology capability.
The same report said nearly one-third of Deloitte's global workforce is already based in India, and that about 30,000 employees had been trained in AI capabilities. These are not opposite strategies. They are the same calculation from different angles: cut or combine where AI makes the old structure too heavy, hire where clients still need people who can make the tools work.
None of this is subtle about who feels it first. India's acceleration centres exist because they are efficient at scale, and repeatable work at scale is precisely where large language models and agentic systems are improving fastest. A city or office built around producing spreadsheets, testing controls and assembling slides more cheaply than someone in New York or London has to offer something more once the software can do the basic work in seconds.
PwC's answer is to move that workforce closer to AI engineering and global client delivery, rather than leave it as a lower-cost execution layer. That bet is plain. Whether 40,000 people can be retrained fast enough is the part no announcement can settle.
If you work in professional services anywhere in the offshore delivery chain, from Bengaluru to Manila, this is the clearest dated signal yet that the big firms are done waiting. The model is changing first. The job titles will follow.
Also read: AI Just Helped Healthtech Startups Mint Twice as Many Unicorns in 2025 • DeepSeek's New AI Model Spooked Samsung and SK Hynix Investors • Sam Altman tells Fortune an OpenAI IPO is off the table for 2026 This article is posted in AI News , check it out for more related stories.
Source: Startup Fortune
