I took your example data and turned it in to reproduceable DML/DDL which is really helpful for these sorts of questions. I added a ProductID column, and a few extra rows of data to demonstrate how to do it with multiple IDs. I used table variables because there's less to clean up, but that's just personal preference.

DECLARE @PriceHistory TABLE (ProductID BIGINT, Price DECIMAL (10,2), Date DATE)
INSERT INTO @PriceHistory (ProductID, Price, Date) VALUES
(1, 20, '2023-01-01'), (1, 20, '2023-01-02'), (1, 20, '2023-01-03'), (1, 19, '2023-01-04'),
(1, 20, '2023-01-05'), (1, 20, '2023-01-06'), (1, 20, '2023-01-07'), (1, 20, '2023-01-08'),
(2, 30, '2023-01-01'), (2, 15, '2023-01-02'), (2, 15, '2023-01-03'), (2, 19, '2023-01-04'),
(2, 19, '2023-01-05'), (2, 20, '2023-01-06'), (2, 20, '2023-01-07'), (2, 20, '2023-01-08');

Using this, we can query using a recursive common table expression to get the results you're looking for. To separate the ranges I used a LAG windowed function to flag if this row belongs to the same range as the previous one.

;WITH base AS (
SELECT ProductID, Price, Date, CASE WHEN LAG(Price,1) OVER (PARTITION BY ProductID ORDER BY Date) = Price THEN 1 ELSE 0 END AS InRange
  FROM @PriceHistory
), Ranges AS (
SELECT ProductID, Price, Date AS StartDate, Date AS EndDate
  FROM base 
 WHERE InRange = 0
UNION ALL
SELECT a.ProductID, a.Price, a.StartDate, r.Date AS EndDate
  FROM Ranges a
    INNER JOIN Base r
      ON a.ProductID = r.ProductID
      AND a.EndDate = DATEADD(DAY,-1,r.Date)
      AND a.Price = r.Price
)

SELECT ProductID, Price, StartDate, MAX(EndDate) AS EndDate, DATEDIFF(DAY,StartDate, MAX(EndDate))+1 AS Duration
  FROM Ranges
 GROUP BY ProductID, Price, StartDate
 ORDER BY Ranges.ProductID, Ranges.StartDate;