Independent
Software Developer
HR Manager
Team Leader
Accountant
Copy Writer
Customer Service
Virtual Assistance
Editor’s note — 1 October 2026: When this column shipped on 30 September, we described Trump’s $103,265 H-1B fee as having “taken effect on September 21.” That isn’t accurate. The September 2026 proclamation extended the fee’s authority through 21 September 2027, but the fee itself has been blocked since July 2026 by a First Circuit court order barring USCIS from collecting it; the government’s request for a stay was denied. The fee is not currently being charged.
The policy mechanism the column built its argument on — an enforced onshore surcharge repricing renewals toward offshore delivery — therefore rests on a cost that is not being paid. The Concentrix Q3 half of the piece (50 percent AI-native revenue, growing 30 percent year on year) stands on the record. The policy half does not.
We regret the error, and we’re grateful to the reader who caught it. The body text has been corrected in place to reflect the fee’s current blocked status. A correction email is going to the subscribers who received the original edition. — The Editor
The Fee That Subsidized Offshoring
On September 21, Trump extended his $103,265 H-1B fee proclamation through September 2027 — but the fee itself has been blocked since July 2026 by a First Circuit court order and USCIS is not collecting it. On September 29, Concentrix printed its third-quarter results, with revenue guiding down and 50 percent of the top line now from AI-native business models growing 30 percent year on year. Read together, the two disclosures describe the same delivery model from opposite ends. What the fee raised, the operator just told the market it can supply — offshore, AI-native, and at a fraction of the loaded onshore price.
Concentrix ended its September 29 call with the number that has now decided the trade.
Fifty percent of the company’s revenue is coming from AI-native business models, and that segment is growing 30 percent year on year.
Management guided to more than $6 billion in new business revenue by 2027, roughly 2.2 times the current book.
The legacy contact-center line, priced against U.S. onshore labor, is running off at 3 to 5 percent a quarter. On the same slope, the crossover is closer than the punished stock price suggests.
That is the operator side of the ledger. The buyer side started nine days earlier.
The Trump administration’s $100,000 H-1B fee was signaled again on September 18, when a presidential proclamation extended the fee’s authority through September 21, 2027. The fee itself has been blocked since July 2026 by a First Circuit court order barring USCIS from collecting it; the government’s request for a stay was denied. The fee is not currently being charged.
Moody’s had estimated that, if the fee were enforced, it would add $100 million to $250 million a year in operating expense across the largest Indian IT firms. Today that number sits in the risk column, not the cost column.
The more consequential number, for the reader running a services P&L, is what the fee WOULD do to the price of onshore U.S. technical work if enforcement reactivated. The current status — proclamation extended, court order blocking collection — leaves it as a signal rather than a line item. But the signal matters at renewal: a vendor pricing a multi-year onshore contract today has to decide whether to hedge against the fee being reinstated, and that hedge is itself upward pressure on quoted rates.
Where the fee actually pushes the work
A Wharton read published by SHRM after the fee took effect names the destination directly.
Studying the last major H-1B tightening cycle, Wharton economist Britta Glennon found that for every rejected H-1B petition, corporations hire between 0.4 and 0.9 additional employees abroad.
In Glennon’s own language, the fee “could push jobs overseas instead of protecting American workers.”
Fwd.us’s policy brief reached the same conclusion from a different angle: the pricing, if applied, would raise the cost of the onshore option without adding capacity anywhere else onshore. The research describes the behavior under enforcement; the current enforcement status is zero, but renewal planning cycles are long enough to price the risk anyway.
Nothing about the mechanism sends the work back to Columbus.
The physical layer that receives it is already built.
India’s Global Capability Center ecosystem crossed 2,000 centers in the past twelve months, employing 1.9 million professionals across roughly 25 percent CAGR.
The Philippines’ IBPAP roadmap, backed by $25 million a year in reskilling funding, targets 2 million AI-enabled workers by 2028.
Neither footprint was scoped for a U.S. policy tailwind. Both now have one.
What Concentrix’s split says about the trade
The Concentrix headline number was not encouraging.
Revenue $2,453.7 million, down 1.2 percent year on year. Q4 guide: revenue $2.4 to $2.5 billion, EPS $2.86 to $2.98, both below consensus.
Management cited hyperscaler spend priorities as the primary Q4 headwind. The stock got punished before the earnings call ended.
The half of the release the market largely skipped is the composition.
AI-native at 50 percent of the top line, growing 30 percent.
The operator that ran a quarter-million seats of contact-center capacity three years ago just told analysts that half its revenue is now services-as-software delivered from lower-cost geographies.
That is also the shape of the buyer’s next renewal quote.
When a chief financial officer weighs an onshore staffing model carrying a $103,265 policy risk against a delivery model already priced offshore and already AI-native, the arithmetic bends the same way: the signal only has to crystallize once for the renewal math to swing.
What the buyer does now
A finance chief comparing renewal options this quarter has three routes forward.
Keep the onshore staffing model and absorb either the hedge against reinstatement or, if the courts flip, the full $103,265 line item. Rebuild the case for the CFO either way.
Move the workload to an offshore third-party operator whose delivery is 40 to 60 percent of loaded onshore cost and whose Q3 print shows the AI already embedded.
Stand up a captive GCC in Bengaluru or Manila and take the delivery in-house, skipping the vendor markup entirely.
Two of the three routes end offshore.
Two complications sit on top of that.
Concentrix flagged hyperscaler spend priorities as its Q4 headwind, which means the AI-native growth rate compresses if enterprise cloud budgets contract, and operators sitting on the wrong side of a hyperscaler shift will not enjoy this trade.
And HFS Research’s 2026 report remains the honest ceiling: only 21 percent of enterprises have booked any revenue from AI so far, and zero percent have reached what HFS calls the agentic-enterprise stage.
The buyer capable of running the offshore-plus-AI trade cleanly is a narrower cohort than the sales pitch suggests.
Neither complication disqualifies the trade.
Both change the question a services buyer should be putting to the vendor at renewal.
Not whether the vendor uses AI, which stopped separating operators a year ago.
Whether the vendor has an offshore delivery footprint sitting alongside a measurable AI-native revenue mix, because the operator without either is defending a book of business priced against a policy signal that could crystallize into a $103,265 line item overnight.
The question for your business
At your next renewal, which side of the $103,265 policy risk is your vendor’s delivery model on — and can they show you AI-native revenue that is actually growing, not just an AI slide in the pitch deck?
